'Bowie' vocoder algorithm: requests for comments

i folks, sorry Ive been busy with other issues for a while. Im back to dsp dev now and working on a new kind of vocoder called 'Bowie.' Below is top-level structure of Bowie phase-banked AV vocoder, now in alpha stage, prior to conversion to gen~ code. By default, Bowie provides up to 32 detuned voices with a common dual quad-band vocoder for each voice; changes in vocoder settings alter the video frame pointer, which in the demo below simply has a Bowie picture from his album 'Fame.'

For the vocoder, I am taking of a new approach: starting for example with a 96kHz audio stream in from a mic, it first filters to 19.2kHz, then upsamples with spline interpolation to 96Khz bitrate, and runs a 1024-point FFT on that to create the four windowed bands, then decimates back to 19.2kHz for output. I believe this would result in better quality and more flexible options than standard approaches, also by putting in a poly~ object with voice-specific controls, could additionally work as a harmonizer. On checking music catalogs, hardware versions of such harmonizers still typically run $1000, so many people could be interested in the final application (and source code, which will be in the gen~ Synthcore v3 library). I hope to make a nice demo app on Windows and MacOs for free, and before I start on detailed coding of the DSP algorithms, I was hoping to hear comments from those in this community familiar with this kind of app.Also it was announced on Memorial Day Weekend: Synthcore v2.1 will remain at its current price of $15 until Tuesday June 12th. With felicitations to Berlin on 2018 Memorial Day weekend.